Research Michigan Center Retirement

advertisement
Michigan
Retirement
Research
University of
Executive Summary
Center
The Welfare Implications
of Increasing DI Benefit Generosity
John Bound, Julie Berry Cullen,
Austin Nichols, and Lucie Schmidt
IB 2002-024
July 2002
M R
R C
Issues in Brief
Project #: UM 00-12
Researchers studying the impact of disability insurance (DI) in
the U.S. have focused on the impact of program parameters on
the costs of the program. What is striking, however, is that there
have been no studies that attempt to place a value on the benefits
associated with providing DI. The effectiveness of the program
depends on how the costs relate to the social benefits. Given its
importance and continuing growth, an evaluation of whether the
DI program provides adequate insurance against the income
losses associated with the onset of severe limitations is overdue.
We address this question by evaluating whether workers are currently over or under-insured against career-ending disability. To
test which is the case, we simulate the welfare implications of a
small increase in the level of payments.
In this Issue in Brief, we summarize a study that evaluates the
potential impact of a marginal increase in DI benefits. Our simulation allows us to assess the potential costs and benefits. The
cost of increasing DI benefits by 1% has two components: the
direct or static cost, and the indirect or dynamic cost. In short, the
direct cost is the cost to the government of the benefit increase
itself, while the indirect cost is the cost to the government of
changes in people’s economic behavior caused by the benefit increase. The extent of these costs and individuals’ valuations of
the net benefits are estimated using data from the March 1991
Current Population Survey (CPS) and the 1990-1993 waves of the
Survey of Income and Program Participation (SIPP). In general,
our results suggest that the level of DI benefits is close to its optimal level.
The Welfare Implications…
1
Summary of Major Findings
Direct and Indirect Costs of the Reform
The most obvious direct cost of a 1% increase in DI payments is the cost of the higher payments themselves to current
DI beneficiaries. We expect the government to pay an extra
$320.7 million in DI benefits to current DI beneficiaries as a result of the benefit increase. However, the total direct cost of the
benefit increase is less than this because the increase in DI benefits reduces the cost of other government programs. Since DI
benefits are income that counts against DI beneficiaries’ eligibility
for Supplemental Security Income (SSI) and Food Stamps, a 1%
increase in DI can be expected to reduce SSI and Food Stamps
payments to current DI beneficiaries by $28.9 million and $0.9
million, respectively.
Furthermore, DI income is taxable, so an increase in DI benefits will increase the revenue the government receives from current DI beneficiaries in taxes. We expect a 1% increase in the
generosity of DI benefits to increase federal tax revenues by $4.1
million and state tax revenues by $0.5 million. After subtracting
the tax, SSI and Food Stamps savings to the government, we estimate the total direct cost to the government of a 1% increase in
the generosity of DI benefits to be $286.3 million.
If the benefit increase had no effect on people’s economic
behavior, the total cost of a 1% increase in DI benefits would
equal its direct cost. But since the higher DI benefits change the
economic incentives people face, the benefit increase has indirect
costs to the government, too. First, higher DI benefit payments
should reduce the work hours, earnings, and tax liabilities of the
family members of DI beneficiaries. This is because higher DI
benefits will increase the income of the families of current DI
beneficiaries, reducing the incentives of family members to work
long hours. Assuming that a 1% increase in the generosity of DI
payments will result in a 1% decrease in the earnings of wives of
male DI beneficiaries but no change in the earnings of husbands
of female DI beneficiaries, we estimate the cost to the
government of the tax revenues lost as a result of this drop in
spousal earnings to be $2.2 million.
Second, an increase in the generosity of DI benefits should
cause more people to apply for DI benefits. Results from other
studies suggest that a 1% increase in the generosity of DI
payments will result in a 0.5% increase in applications for DI
benefits. Analysis of data from the 1990-93 SIPP and from Social
Security disability determination records suggests that the
proportion of these extra applicants who are awarded DI benefits
will be approximately equal to the proportion of all applicants
who are awarded DI benefits. Consequently, we expect a 0.5%
increase in DI applications to result in a 0.5% increase in DI
beneficiaries. We estimate the cost to the government of paying
DI benefits to these extra beneficiaries to be $89.7 million.
These extra DI beneficiaries will face work restrictions as a
condition of receiving DI benefits. Indeed, the vast majority of
DI beneficiaries do not work at all. Therefore, these extra DI
beneficiaries will work fewer hours, earn less income, and pay
fewer taxes than before. We estimate that these lost tax revenues
will cost the government an additional $50.9 million. With a
general net loss of income as a result of reducing work hours or
The Welfare Implications…
2
leaving work altogether to receive DI, these extra DI beneficiaries
will become more eligible for Aid to Families with Dependent
Children (AFDC), SSI, and Food Stamps. We expect the cost of
these extra DI beneficiaries through these programs to be an
additional $0.2 million. As a result, the total cost to the
government of extra beneficiaries attracted to DI by the benefit
increase is $140.8 million.
Between the attraction of extra beneficiaries to the DI
program and the reduction in work hours, earnings, and tax
liability of the spouses of DI recipients, the total indirect cost of a
1% increase in the generosity of DI payments is $143.0 million.
This, along with the $286.3 million in direct cost, adds up to a
total cost of the benefit increase of $429.3 million. Note that the
total cost is 1.5 times the direct cost. This implies that increasing
the incomes of current DI beneficiaries by $1 has a cost to the
government of $1.50--$1 to increase the beneficiaries’ incomes,
$0.50 as a result of the side effects.
Representative Workers’ Willingness to Pay
On its face, this analysis suggests that the DI program, or at
least expanding the DI program, is inefficient and undesirable.
However, this is not necessarily the case because people are risk
averse. In the absence of DI, people face dramatically lower incomes in the event of being disabled than in the event of not being disabled. As a result, the risk of disability to most people is a
risk of having to survive on a considerably lower income than
that which they otherwise would have earned. DI reduces this
financial aspect of the risk of disability by acting as an insurance
program that redistributes income from the event of not being
disabled to the event of being disabled. Because it reduces the
risk of a severe reduction in income due to disability, people
might support expanding DI even if the expansion is inefficient.
The degree to which people will support expanding DI despite
inefficiency depends on the degree to which they are risk averse.
If they are not very risk averse, they do not care much about the
risk of a lower income due to disability, and consequently will not
tolerate reducing this risk inefficiently by expanding DI. On the
other hand, if they are very risk averse, they will tolerate great inefficiencies in reducing this risk by expanding DI.
Is reducing the loss of income in the event disability worth the
inefficiencies of expanding DI? In general, the answer appears to
be yes. From our analysis using data from the 1990-93 SIPP, we
expect that, given reasonable assumptions about the degree to
which people are risk averse, the average worker would favor a $1
increase in net disability benefits even if the cost to the government—and, eventually, to themselves through taxes—is greater
than $1.50.
Enthusiasm for DI expansion may differ across people with
different levels of education and different incomes. People with
more education and higher incomes, for example, may be more
inclined to support more generous DI benefits because they face
a greater loss of income in the event of becoming disabled and
would favor greater DI benefits to help offset that loss. On the
other hand, people with more education and higher incomes also
pay a larger share of the tax burden of DI benefits and are less
likely to become disabled and collect DI benefits. For those reasons, people with more education and higher incomes may be
The Welfare Implications…
3
less inclined to support higher DI benefits, considering only their
own financial outcomes.
Analysis using data from the March 1991 CPS and the 199093 SIPP suggest that the latter effect is dominant and that selfinterested support for DI expansion should be lower among
more educated people. The average person with a high school
diploma and no college faces a 3.6% probability of becoming disabled. If the generosity of DI payments is increased by 1%, he
can expect to pay an extra $3.40 in net taxes in the event of not
being disabled and receive an extra $62.96 in net benefits in the
event of being disabled. Given this information and assuming a
reasonable degree of risk aversion, the average person with a high
school diploma and no college would gain from the 1% benefit
increase. On the other hand, the average college graduate faces
only a 1.2% probability of becoming disabled. A 1% DI benefits
increase raises his net taxes in the event of not being disabled by
$5.88 and his net benefits in the event of becoming disabled by
$68.53. Given a reasonable level of risk aversion, the average college graduate would not gain from an expansion of DI benefits at
this cost. More generally, our results show that DI is not a particularly “good deal” for the typical person with more than a high
school education.
Life Cycle Incidence for Average Workers
So far, we have only considered how typical workers might be
affected by an expansion of DI benefits with regard to the risk of
becoming disabled. However, people face many risks other than
the risk of becoming disabled, including the risk of becoming
poor for reasons other than disability. An increase in the gener-
osity of DI benefits will make the non-disabled poor worse off
because the benefit increase will increase the tax burden, however
small, on them. If the non-disabled poor are worse off than most
of the disabled, individuals may oppose expanding DI and redistributing income from the former to the latter. Consequently,
risk aversion might cause people to oppose increasing DI benefits.
Indeed, our analysis suggests that this is likely to be the case. Using the March 1991 CPS data, we find that, given a consideration
of all economic risks, including those of becoming disabled or
simply poor, people are made worse off by a 1% increase in the
generosity of DI benefits. The reason is that DI redistributes income to people in bad economic states (the disabled) from people in even worse economic states (the non-disabled poor).
Incorporating Health Insurance
We have also not considered health insurance and medical expenses in our analysis so far. DI beneficiaries automatically receive Medicare health insurance coverage. As a result, when the
benefit increase attracts extra beneficiaries to the DI program, it
increases the number of Medicare beneficiaries, too. The cost to
the government of these extra Medicare beneficiaries adds $45.3
million to the indirect cost of a 1% increase in the generosity of
DI benefits. With this extra indirect cost, the total cost to the
government of increasing the income of DI beneficiaries by $1
increases from $1.50 to $1.66. This change does not substantially
affect our conclusions about whether typical workers would be
willing to trade lower take-home pay for higher DI benefits.
However, considering medical expenses does notably affect
our conclusions about how exposure to other risks affects the
The Welfare Implications…
4
attractiveness of a benefit increase. Despite Medicare coverage,
the disabled have much higher out-of-pocket medical expenses
than the non-disabled. As a result, the disabled are not as much
better off in terms of disposable income than the non-disabled
poor as our initial analysis implied. In this light, people are not as
likely to take the risk of becoming one of the non-disabled poor
much more gravely than the risk of becoming one of the disabled,
and so it becomes less important to them that DI redistributes
income from the former to the latter. Repeating our analysis with
health insurance and medical expenses included in consideration,
we find that people are, on average, made only slightly worse off
by a 1% increase in the generosity of DI benefits. The sensitivity
of our results to relatively small changes in the analysis suggests
that the level of DI benefits is not too far from the optimal level.
Conclusion
In this brief, we consider the welfare implications of increasing the generosity of increasing Disability Insurance (DI) benefits.
We use data from the March 1991 Current Population Survey
(CPS) and the 1990-93 waves of the Survey of Income and Program Participation (SIPP). We find that the cost to the government of increasing the after-tax incomes of DI recipients by $1 is
$1.50 due to the effects that the benefit increase will have on the
behavior of DI beneficiaries and their families and of potential
DI applicants. This figure increases to $1.66 when we include
health insurance and medical expenses in our analysis. Due to
risk aversion, we find that typical workers are made better off by
an increase in DI benefits despite this inefficiency because DI reduces the risk of having to survive on a low income in the event
of becoming disabled. However, when other risks, such as the
risk of becoming poor but not disabled, are included in our analysis, we find that people are not made better off by an increase in
DI benefits. Our results do suggest that the level of DI benefits
is close to its optimal level.
The Welfare Implications…
5
John Bound is a Professor of Economics at the University of
Michigan.
Julie Berry Cullen is an Assistant Professor of Economics at the
University of Michigan.
Austin Nichols is at the University of Michigan.
Lucie Schmidt is at the University of Michigan.
This work was supported by a grant from the Social Security Administration through the Michigan Retirement Research Center
(Grant # 10-P-98358-5). The opinions and conclusions are solely
those of the authors and should not be considered as representing the opinions or policy of the Social Security Administration or
any agency of the Federal Government.
Michigan Retirement Research Center
Institute for Social Research
University of Michigan
426 Thompson Street, Room 3026
Ann Arbor, MI 48104-2321
Phone (734) 615-0422
Fax (734) 615-2180
http://www.mrrc.isr.umich.edu
mrrc@umich.edu
Regents of the University of Michigan
David A. Brandon, Ann Arbor
Laurence B. Deitch, Bingham Farms
Daniel D. Horning, Grand Haven
Olivia P. Maynard, Goodrich
Rebecca McGowan, Ann Arbor
Andrea Fischer Newman, Ann Arbor
S. Martin Taylor, Grosse Pointe Farms
Katherine E. White, Ann Arbor
Mary Sue Coleman, ex officio
The Michigan Retirement Research Center is supported by a grant
from the Social Security Administration (grant number 10-P-98358-5).
The Welfare Implications…
6
Download